Data Sampling
- Non-probabilistic sampling
- Simple random sampling
- Stratified sampling
- Weighted sampling
- Importance sampling
- Reservoir sampling
- especially useful for streaming data
- The algorithm involves a reservoir, which can be an array, and consists of three steps:
- Put the first k elements into the reservoir.
- For each incoming nth element, generate a random number i such that 1 ≤ i ≤ n.
- If 1 ≤ i ≤ k: replace the ith element in the reservoir with the nth element. Else, do nothing.
